6.1 Water technology
6.2 The capacity and density of water points
6.3 Organisation and management
While the scope of this study includes all parts of tropical Africa, and all the livestock production systems therein, by far the greatest part of the experience gained in the past, or at least of that which has been recorded and analysed in publications and reports, was gained in arid and semiarid regions. This is not surprising in that it is in those regions that the physiological demand for water is greatest, that most of tropical Africa's domestic livestock live, from which their output comes, and where the supply of water is most limited. It does, however, make it difficult to prescribe for the future development of other regions on the basis of past experience. Most prescriptions that follow apply therefore primarily to the dry regions.
Much of the discussion and planning of water development in the past has revolved around the optimum location and spacing of water points and their capacity, in terms of the number of livestock they can water. The focus has been on the objectives of increasing livestock output and avoiding environmental degradation. However from the present study, several other points have emerged as requiring at least equal attention in the future. Some of these points have causal linkages between them. One of the points is that equity considerations - who benefits or suffers from water development - should rank at least equal with the other objectives. Another is that the effect of a new water supply in reducing the amount of labour required for watering is as important to some livestock owners as its effect in reducing the distance livestock trek to water. A third point is that the failure and breakdown rates of water points are high - too high for sophisticated calculations about optimum location to be valuable unless they specifically take these high failure rates into account. Technical, financial and administrative reasons all lie behind the failure rates and each needs to be given specific attention. Systems and procedures of control and management both interact with technology and have independent effects of their own. Finally, water development is often attended by unexpected and unintended changes in land use. This chapter provides some pointers/to future development in respect of water technology and its organisation and management, as well as of land use and location and spacing of water points.
The decision to develop water supplies in an area involves not only deciding how many supplies there should be and where they should be located, but also what the source of water should be and how it should be extracted from the source and distributed to livestock. In some areas technical factors, e.g. rainfall, run-off coefficients and soil porosity which affect the viability of dams and hafirs, or geology which affects the prospects of finding underground water, strongly influence the choice in one direction or another. In other areas the options are more evenly balanced. Even where for technical reasons one kind of source, i.e. underground or surface, is strongly indicated, there are options about how the water point is to be constructed and how water is to be extracted from the water source and distributed to livestock. Up to depths of 100 m it is possible to dig open wells by hand labour instead of drilling a borehole, and it is possible to raise water from this depth by human or animal rather than by mechanical power. Similarly it is possible to build dams and to excavate hafirs - and to clean them of silt from time to time - by human and animal rather than by mechanical power. In the same way options also exist between different kinds of drilling devices, between different kinds of mechanical power and between different combinations of human labour, animal power and machinery or tools. Options exist - and the choice between them needs to be based on technical, economic, environmental and social considerations, not just on one of these alone.
An important element in the choice to be made is the requirement for human labour. Water extraction from open wells by human power is extremely labour intensive, requiring up to twice the labour force at peak times compared to some other extraction methods (Swift, 1979, Ch. 5), and animals waste much time at well heads waiting for water. On the one hand if the incomes of both livestock owners and herders are to be substantially improved in the long run, increases in labour productivity will be required. On the other hand high labour requirements for watering are the most effective limit currently available on the size of the livestock population; and unless alternative employment opportunities can be found for those displaced by less labour-intensive watering techniques, the effect of these techniques, rather than to increase aggregate income, will be to reduce the drudgery of watering, to increase leisure time and to redistribute income in favour of those with large herds and little family labour at the expense of those with labour but not enough livestock.
In principle the argument about labour-intensive water extraction is the same in all zones, for all production systems and for all species of livestock. In practice it is with the large herds of nomads and seminomads in the arid and semi-arid zones that it is of significant importance. There is a little evidence (Swift, 1979, Ch. 5) that watering camels from wells leads to sharper relative peaks in labour requirements than in the case of other livestock. Much labour can also be used in constructing water supplies, especially dams and hafirs. In this case the arguments probably indicate that more emphasis should be given to labour-intensive techniques of construction in the high-rainfall zones and their associated production systems, partly because the productivity of manual labour in such zones tends to be higher than among the pastoralists of dry regions, and partly because it is much easier in the more densely populated high-rainfall zones to collect a labour force of adequate size to be effective without totally disrupting family life and economy. To the extent that requirements for male labour in the dry season are particularly high for herding camels it is especially difficult to raise a labour force big enough to be useful for constructing a water supply from a camel-herding system.
The very high breakdown and failure rates for water supplies constructed under livestock development programmes in dry zones indicate the need and scope for substantial improvements in technical efficiency. However, it should be noted that all components, not just water development, have a very high failure rate in livestock programmes in dry regions (which indicates a particularly difficult physical and social environment) and that comparable figures for water supplies in other zones are not available, so it is not possible to assess, on the basis of firm data, to what extent the problem is peculiar to these dry zones. General impressions are that it is, and the known greater climatic variability of the dry regions over the more humid would indicate that, for surface water resources at least, more problems are likely to occur due to miscalculation about dam and spillway capacities. The low technical efficiency stems from inappropriate technology, inadequate training and information, and defective organisations and management, including financial and administrative procedures.
Given the high breakdown rate, priority in choosing the technology of water development should be given to answering the questions 'What can be done when something goes wrong?' and 'How will this water point be serviced and maintained?' All rural regions in Africa have acute problems with the availability of technical skills, with stocks of spare parts and consumables (e.g. diesel fuel) and with transport systems; the dry rural regions in Africa suffer these problems to an extreme degree. The choice of technology should not only be determined - as it largely is now - by which form of initial construction is cheapest, quickest and can be most easily financed (e.g. from foreign aid sources), but also by which technology can be kept going with the resources of the local area where it is based. These considerations are likely to indicate that open wells, even very deep ones, should much more often form part of water development programmes than they do now.
Part of the reason for the high failure rate is lack of adequate personal experience by design and site engineers and locally tested and adapted model designs. For example, in southern Ethiopia in the 1980s dams and hafirs were being modelled on those featured in an Australian publication because there was no African - far less an Ethiopian one - available. Not only are relative skill endowments and prices likely to differ systematically between Australia and Africa but soils, climate and other physical factors differ as well.
Experience will come with time and the problem of African-adapted designs could relatively easily be solved. But there is also a lack of site-specific information on hydrology and hydro-geology, and this will not be so easily solved, both because of the cost of acquiring this information and because, in many cases, it is only when it has been collected over many years that it yields useful results, e.g. on peak flood levels. In some cases inadequate efforts are made to use or collect information which is available in the memory of local inhabitants, and a greater willingness by design engineers to collect this information is required. This will increase the length of time it takes to design individual water points. But given the present high failure rate, it seems likely to lead to more water points that actually operate.
More attention needs to be paid to training local people in the area concerned in construction, operation and maintenance of water points. Often water points break down because of simple mistakes made out of ignorance; often technicians (and machinery) have to be brought in over huge distances and at great expense, either to do very simple jobs (e.g. replace a washer, mend a crack with some cement) for which a minimum of equipment, training and confidence are required, or to repair equipment that is unnecessarily complex in the first place. Often it is not formal training that is needed but familiarity acquired from being allowed to handle and strip down equipment (Hitchcock, 1979, p. 171).
It is not only the inefficiency of over-complex equipment or of the absence of training for local people on how to do repairs themselves which is important. It is also the fact that it makes local production systems so dependent on outside help, help which may demand onerous terms and whose income will not then be recirculated within the area to foster the growth of a locally diverse economy. Local technicians or craftsmen may exploit their neighbours - but at least the incomes gained from this exploitation will, in part at any rate, be locally spent. Complex technology requires highly skilled technicians who command high rates of pay. Sometimes even when technicians with the requisite skill can be recruited from local sources this will not be done because the rewards of the job are set by political, not by market forces, and the jobs represent prizes to be given to those economic or ethnic groups most closely related to central government's political support. Less complex technology does not require skills so highly rewarded as to form part of government's patronage25
25 This point can usually be demonstrated by looking at the origins of vehicle drivers and unskilled labourers in government pastoral development projects. Unskilled labourers usually come from the local area's ethnic group(s). Drivers, even though there are unemployed drivers from local ethnic groups available who want employment, usually come from ethnic groups closer to central government.
There is an equity element about the choice of technology which needs to be kept in mind, and these problems are most likely to arise in arid and semi-arid zones which are often both politically and geographically marginal to the nation state that incorporates them.
The capacity of water points - expressed in terms of the volume of water they can supply in a given period of time - and their density - which we can express in terms either of water points per km2 or in the distance (km) between water points - will influence the species, breed and age/sex composition of the herds, as well as the yield of useful products per animal and the nature and intensity of pressure exerted on the soil and vegetation. This section considers issues of capacity and density in relation to the different production systems distinguished in Chapter 1.
In high-rainfall areas closely spaced water supplies are a precondition for the emergence of a smallholder dairy system, both because of the disease problem and because plentiful water is needed to allow high-yielding dairy stock to express their genetic potential. Where good markets for fresh milk exist and water supplies are adequate in number, cattle will form a high proportion of the total livestock composite, and cows of breeding age, probably containing a high proportion of exotic blood (often Friesian), will constitute a large part of the cattle population - as much as 50% (Goldson, 1980).
In highly productive smallholder dairy systems water supplies need to be sufficiently numerous that disease-susceptible stock of high value do not need to leave the boundaries of their farm to water at places where they may be exposed to infectious disease, parasites (especially liver fluke) or injury. This may justify the provision of water points - which may be piped supplies - at a density as high as one point to 10 ha or less. Where the disease problem is not paramount, the high productivity of selected dairy animals may, nevertheless, justify, in terms of energy and time saved in trekking to water, a spacing between water points of as little as 5 km, so that no animal has to trek more than 2.5 km to water. Table 5 in Chapter 4 shows that a 5 km spacing between water points could lead to yields per dairy cow 5% higher than a spacing of 10 km. For a dairy cow yielding on average 2000 litres per year at US$ 0.30 (at farm gate prices) per litre, that represents a US$ 60 increase (gross before deducting the expenses of increased water supplies) in value of milk output per cow per year26. That would justify, on the basis of a 12% interest and 15 years amortisation, an investment of up to about US$ 26000 per additional supply if the cattle population of high-yielding dairy breeds is about 10 per km2 the Kenyan highland average in the 1970s27. A further increase in density reducing spacing between water points from 5 km to 2.5 km would justify an investment per additional water point of up to only US$ 2600.
26 These figures are compatible with data contained in ILCA (1981) with some allowance for inflation in prices. The calculations of Table 5 are based on particular assumptions which may not be realistic for smallholder dairy systems.27 Assumes a cow: follower ratio of 1:1; this is compatible with data in Goldson (1980).
We do not possess adequate evidence to make the same sort of calculations for mixed farming systems in high-altitude areas. Here the most valuable form of economic output may be draught power from oxen. It is difficult both to quantify the effect of more plentiful water supplies on this and also to give it a unit value, since it is not a final output for consumption with a market value but a factor of production whose marginal product it is difficult to estimate. In the case of both smallholder dairy and mixed farming systems in high-altitude areas, in both of which the individual herd size tends to be low, an alternative to building additional water supplies is to transport water e.g. on donkey back, from the water source to the livestock which need it. An additional element to be taken into consideration in both smallholder dairy and mixed farming systems in the highlands is the impact on soil erosion of large numbers of livestock trekking to water points. The livestock population density in the highlands of Africa, at an average of about 20 to 25 livestock units (250 kg) per km2 (ILCA, 1981), is 3 to 5 times that of the arid and semi-arid areas, and the topography of the highlands, their relatively high and intense rainfall, and in many cases also their soils, make them inherently more erodible. If water points in the highlands are at 5 km spacing this implies an average of 500 livestock units per water point. With cattle having to trek along narrow paths between fields, that is a sufficient number to cause serious erosion. In practice in highland areas water sources are normally much denser than the 0.05 and 0.20 per km2 implied by these trekking radii of 2.5 and 1.25 km. For example, in the central highlands of Ethiopia the average water point density is 0.87 per km2 (of which flowing streams and rivers account for 0.57), and in only 20% of this whole area of 100000 km2 does the overall density fall below 0.2 (Watson, 1973b, p. 23 and Table 8A).
In the arid and semi-arid zones the factors affecting the appropriate density of water supplies are probably more numerous and more complex than in the highlands. One factor is that in many semi-arid areas where crop cultivation exists and where cultivators and herdsman are from different ethnic groups, the damage done to crops by livestock on their way to water, and the way in which access by livestock to water is barred by the position of cultivated fields, is an important source of inter-ethnic conflict. In these areas, although the amount of labour required to water livestock may be an important constraint on livestock numbers, the actual number of water points is probably not. Extra water points may be justified in order to reduce conflict rather than in terms of extra production. They may cause environmental degradation, but this is more likely to be as a result of the extra spread of cultivation that they permit rather than the increase in livestock numbers.
In arid areas water points are important not only in terms of the grazing which they permit in their vicinity, but also as transit points for animals on migration from one general grazing area to another. Such permanent and reliable transit water points can be extremely important in permitting the movement of livestock (in pastoralists' breeding herds as well as in the herds of traders), especially in times of drought, from poor to good grazing areas. In the absence of such water points herds can be cut off and suffer great losses. Along migration corridors of this kind reliable water points at intervals of about 20-30 km are suggested28
28 Lewis (1978) reports 'weaker members' (the beedi) of transhumant cattle herds trekking 50 km without water and 'stronger' (garti) 90 km. But, especially in a drought year, very few waterless stages of this extent could be completed without huge losses.
Where other methods of controlling livestock numbers (e.g. government regulation, voluntary decision or agreement of livestock owners) are ineffective, the spacing and capacity of water points can be used as an alternative control. Some areas are unsuitable for dry-season grazing, either because only ephemeral vegetation grows there or because hydrological and geological conditions make the provision of water prohibitively expensive. In such areas hafirs and dams can be built to provide a temporary source of water during, and for a short period after, the rainy season. In a number of countries (e.g. Ethiopia, Kenya) such points have been designed so that the volume of water conserved there is no more than enough to water livestock for whom the area's feed sources are sufficient without risk of overgrazing.
The theory of this device for controlling grazing pressure is attractive; but in practice it is difficult to implement. Firstly both the yield of annual grasses and the quantity of stored water available for drinking each year fluctuate with variations in annual rainfall. However, they may not fluctuate in close proportion to each other, especially in the light of the effects of evaporation and seepage; either water or forage may be inadequate in relation to the other. Secondly rainfall in arid Africa tends to vary tremendously within seasons over quite short distances. The water catchment that feeds a hafir or dam is unlikely to be coterminous with the grazing area that is served by it; the rainstorm which makes the grass grow may not fill the dam, and vice versa. Thirdly dams and hafirs silt up over time; the right capacity just after construction will be too small 5 years later; and so on. In principle this is a good system but any kind of precision in calculation is inappropriate because of these uncertainties.
It has to be accepted that in most years either the water points will be too few and too small and not all the available forage will be used, or they will be too large and too many and some overgrazing will occur. Also, in the case of annual grasses, overgrazing is much more serious during the growing period before seed has formed than later in the dry season; with perennials it is the other way round (Breman et al, 1979/80), so that a further element of uncertainty is introduced as to exactly when the livestock will have access to this grazing. Different grazing pressures, and so different watering capacities, will only be appropriate for a particular season of use. A final problem is that there are economies of scale in constructing hafirs and dams. Normally, the bigger the capacity the lower is the cost per unit of capacity. It is extremely difficult to persuade water engineers of the wisdom of constructing many high-cost low-volume stock ponds when the technical possibilities exist for larger ones which can be used for a longer period each year. The moment the range planner's back is turned the water engineer will dig a deeper longer-lasting pond because, by his criteria, it is more efficient to do so (Shepherd, 1981, pp. 7-8). The pastoralist will normally support the engineer.
Where other methods of regulating livestock numbers are ineffective, the ultimate control over them in arid zones is exercised by lack of water or of forage within range of water at the height of the dry season in a drought year; at this time livestock die of starvation or of thirst or of exhaustion from seeking feed and water. In principle it would be far less damaging to the vegetation if livestock were to die of thirst rather than from starvation after they have grazed the range bare. Again, in practice, it is exceedingly difficult to design water supplies which match exactly the availability of water with the amount of forage which can be safely grazed. The main reasons for these difficulties in the case of hafirs and dams have already been discussed in relation to wet-season grazing areas. Many of the same arguments apply also to wells and boreholes, although the availability of water at the latter is less dependent on current rainfall in the area. It never seems to be possible to restrict the quantity of water supplied by formal administrative orders, i.e. to supply only enough water for a given number of livestock even though the physical capacity exists at that time to supply more. Intimidation of the water point operator by livestock owners and overruling of technical departments by higher political authorities always occur when livestock start dying.
Once more there are great difficulties in regulating the volume of water supplies by physical limitations or the capacity of the equipment. If one believes that the safe grazing capacity of an area is such that only 480 animals should be watered every day, then it is foolish to arrange for the physical capacity of the equipment to be only enough to deliver water for 20 animals an hour (20 animals x 24 hours = 480) with no storage reservoir. The pumps will be under too much strain and there will be no flexibility and room for manoeuvre in the event of mechanical breakdown. If, on the other hand, the equipment is sufficient in capacity to water the safe number of animals with only 6 or 10 hours' operation each day, then there will be irresistable political pressures to operate for more hours each day, and so to water more than the safe number of animals in an "emergency", i.e. when livestock owners want to bring in more animals than the area served by the water point can safely sustain. In fact the most effective physical constraint on capacity is when hard human labour by the livestock owners is required to extract water from its source. Some increase in physical capacity can be achieved by people working longer and harder, but the direct cost to the livestock owners, in terms of physical effort, is such that they re-evaluate the desirability of keeping so large a herd. The result may, however, be very inequitable between different households.
These difficulties in controlling livestock numbers by regulating the physical capacity of water points have directed attention towards regulating the number of water points and their consequent spatial density. As a consequence one often comes across recommendations or policies that permanent water points in arid areas should not be less (more) far apart than, for example, 8 km in Botswana (Hitchcock, 1979, p. 178), 10 km in Sudan (Shepherd, 1981, p. 16), 20 km (Bernus, 1977, p. 54) or even 50 km or more (Marty, 1972, p. 97) in Niger. The appropriate spacing of permanent dry -season water points will partly depend on whether one is mainly interested in economic output or environmental protection. It will also depend on the kind of output (milk or meat) in which one is most interested, the species (camels, cattle, sheep or goats) which can be kept in the area, the breeds (Bos indicus, Bos taurus), and on whether the livestock are herded or free-ranging.
Where animal numbers are otherwise uncontrolled (and the range is not fenced into paddocks), regulating the density of water points will not materially affect overgrazing close to the water point, only further away. Whether under free-ranging or herded management the-first 0.5 to 1 km around the water point will be severely overgrazed. Where camels are kept, grazing may occur up to 100 km away from water, even in the dry season; with cattle, sheep and goats that are in milk the range for herded animals is normally up to about 15 km from water, and for 'dry' animals up to about 30 km, although in drought conditions they will go as far as 50 km (Asad, 1970, p. 28). With all livestock grazing pressure will decrease with distance from water, but with herded animals the gradient of the decline is likely to be less than with free-ranging animals. In hot conditions Zebu cattle will probably be able to range further than Bos taurus breeds.
Given this set of interacting variables there can be no uniquely correct spacing of dry-season water points. If very high priority is given to environmental conservation, the water points will have to be spaced 100 km or more apart; then if the area is otherwise unsuitable for camels large portions of it will be only very lightly grazed by other domestic livestock; but there will be a corresponding large reduction in economic output. If camels are present then even in outer rings around each water point there will be notable grazing pressure; but this should be compensated by greater output. If a large proportion of the herds is females in milk - because males are slaughtered young or sold off early - then total cattle numbers are likely to be kept at a lower level because of the milking cows' inability to range so far. Restricting the density and number of water points is a very costly method, in terms of lost output, of restricting livestock numbers and so grazing pressure.
In future, for reasons of both efficiency and equity, more attention needs to be given to questions of organisation and management than has been given in the past. In this section we deal with three points of paramount importance: the control of access, the ownership of water points, and financial and administrative procedures.
In the arid zone, and to a lesser extent in the semi-arid zone, where water points are few and far between and communal systems of land tenure often prevail, control over access to water is tantamount to control over access to grazing land. The development of new water points is not, therefore, simply a technical matter but also a political one concerning questions of equity to different groups. This needs to be recognized at the outset of a water development programme so that decisions can be made on appropriate political grounds. The criteria for eligibility to access to a new water point, and the adjudication process by which eligibility is decided and enforced, need to be clearly laid down.
The question of ownership of new water points affects partly rights of access, partly the way in which management decisions are made and implemented, and partly financial and administrative procedures. For example, it is possible for a government or the local community to lay down rules on who may or who may not have access to privately owned water points in a certain area, and under what conditions, but in practice it may be infeasible to impose these rules on someone who has already invested his own labour or capital in building a water point in a remote area and who wishes to recoup his money from as many well-off people as possible. It will be difficult enough even to regulate the season of the year in which privately owned water points may be used. In areas where water points are plentiful, usually high rainfall areas or along large sandy rivers where water can easily be found by digging in the bed, it is unlikely that the owners of private wells will be able to reap monopoly profits from their control of a scarce resource or from the access to grazing that this control determines. In areas of dense, settled population land is usually under some form of individual tenure, and the closure of water points by regulation is not an appropriate tool to enforce rotational grazing. In these two kinds of areas private ownership of water points does not carry heavy drawbacks and often has the great advantage of a decentralised form of management and the facility for procuring goods and services for repair and maintenance without being strangled by red tape in the way government departments often are. It may, therefore, be the most appropriate form of ownership in these cases. However, a survey of private water points in the Central District of Botswana (Hitchcock, 1978, Ch. 7) revealed that nearly half are owned by people who own more than one water source and about 99% had absentee owners; in these cases some of the defects of 'centralised' management must be in evidence.
In areas where water points are very scarce and where they provide a useful tool for public management of grazing and access to grazing, private ownership may have severe disadvantages and some form of public ownership, either by local community or by government, may be more appropriate. Community ownership - implying control of access to a water point by a particular social group, possibly a village or a kinship group, and some fairly participatory form of management may be the best arrangement in areas where the human and livestock populations are fairly sedentary. The same may be true even where populations are nomadic, provided that their movements are predictable and regular from year to year and that there is no significant intermingling of different communities at the same watering place (comprising one or several water points within a short distance of each other). If however, nomadic movements are highly irregular and unpredictable, then participatory forms of management of permanent water points become unviable due to lack of continuity of decision-making or of commitment to the long-term success of that point. If different communities water at the same point, or within a very short distance of each other, inter-community competition will probably need to be regulated by government interference. Community ownership of water points is also unlikely to lead to equitable access to water in societies where power and other resources are unequally shared, and government ownership may, in such cases, be less inequitable than ownership by the community.
Where government ownership of water points is required, for one of the reasons already given, two procedures need to be followed. The first is to provide a system whereby the information which local people already have, about hydrology, about the location of preferred vegetation at different times of year, about migration patterns in good and bad years, about the location, size and type of facilities required, can be incorporated into a government's planning processes. The second is to establish financial and procurement procedures suitable for the operation and maintenance of water supplies in remote areas. At the moment few, if any, countries in Africa have these. It is beyond the scope of this study to analyse or specify the matters in detail; but they are a major cause of current inefficiency, and rectification is a precondition to improving the supply of water to livestock in much of tropical Africa. The problems are particularly acute in arid and semi-arid zones where the low population density, long distances, and low political influence of the users of government water supplies make the conventional government procedures least appropriate.
